A low-power array multiplier using separated multiplication technique

Abstract

The authors propose a separated multiplication technique that can be used in digital image signal processing such as finite impulse response (FIR) filters to reduce the power dissipation. Since the 2-D image data have high spatial redundancy, such that the higher bits of input pixels are hardly changed, the redundant multiplication of higher bits is avoided by separating multiplication into higher and lower parts. The calculated values of the higher bits are stored in memory cells, caches, such that they can be reused when a cache hit occurs. Therefore, the dynamic power is reduced by about 14% in multipliers by using the proposed separated multiplication technique (SMT) and in a 1-D 4-tap FIR filter by about 10%.